Abstract
BACKGROUND: Ultrasound-guided fine-needle aspiration cytology (FNAC) is the standard method for evaluating thyroid nodules, but 20–40% of specimens are still cytologically indeterminate (Bethesda III–IV), leading to repeat biopsies or diagnostic lobectomy. We hypothesized that a multiplex droplet digital PCR (ddPCR) assay covering key thyroid-cancer driver mutations would increase the pre-operative diagnostic yield. METHODS: A six-gene ddPCR panel was developed to detect BRAF p.V600E, TERT promoter mutations c.-124 C > T (C228T) and c.-146 C > T (C250T), PIK3CA p.H1047R/L, and all clinically relevant variants in NRAS, HRAS, and KRAS at codons 12, 13, and 61. Analytical performance was established using plasmid controls. The assays were subsequently applied to 210 prospectively collected fine-needle aspiration (FNA) specimens from 201 patients, with parallel ARMS-PCR testing for BRAF p.V600E and histopathological confirmation available for 49 surgically excised nodules. RESULTS: The six-gene ddPCR panel demonstrated 100% analytical specificity and a limit of detection of 0.1% variant allele frequency (VAF). In clinical samples, 103 of 210 nodules (49%) harbored at least one mutation, with BRAF p.V600E being the most prevalent alteration (81/103, 78.6%). ddPCR identified five low-abundance BRAF p.V600E mutations (VAF 0.37–2.96%) that were missed by ARMS-PCR, one of which was confirmed as papillary thyroid carcinoma on postoperative pathology. Mutation prevalence increased progressively across Bethesda categories, ranging from 33% in Bethesda I to 89% in Bethesda VI. Compared with surgical pathology, the combination of the six-gene ddPCR panel and FNAC achieved 100% sensitivity, 71% specificity, and 96% diagnostic accuracy, significantly outperforming FNAC alone or FNAC combined with ARMS-PCR (P < 0.05). CONCLUSIONS: This study developed a six-gene mutation detection panel composed of four multiplex droplet digital PCR (ddPCR) assays for thyroid nodule analysis. Integration with cytology substantially reduces indeterminate diagnoses and may guide personalized management, including the avoidance of unnecessary surgery. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13044-026-00292-9.